This paper considers the creation of Complex Engineered Systems (CESs) and the Systems Engineering approach by which they are designed. The changing nature of the challenges facing Systems Engineering is discussed, with particular focus on the increasing complexity of modern systems. It is argued that modern complexity poses a major challenge to our ability to achieve successful systems and that this complexity must be understood, predicted and measured if we are to engineer systems confidently. We acknowledge previous work which concluded that, in complex systems, failures ("accidents") may be inevitable and unavoidable. To further explore potential tools for increasing our confidence in complex systems, we review research in the field of Complexity Theory to seek potentially useful approaches and measures and find ourselves particularly interested in the potential usefulness of relationships between the magnitudes of events and their frequency of occurrence. Complexity Theory is found to have characterized naturally occurring systems and to potentially be the source of profitable application to the systems engineering challenge, viz., the creation of complex engineered systems. We are left with the tentative conclusion that truly complex systems, with our present understanding of complex behavior, cannot be designed with a degree of confidence that is acceptable given our current expectations. We recommend that the discipline of systems engineering must investigate this issue as a matter of priority and urgency and seek to develop approaches to respond to the challenge.

No abstract

This paper considers the creation of Complex Engineered Systems (CESs) and the Systems Engineering approach by which they are designed. The changing nature of the challenges facing Systems Engineering is discussed, with particular focus on the increasing complexity of modern systems. It is argued that modern complexity poses a major challenge to our ability to achieve successful systems and that this complexity must be understood, predicted and measured if we are to engineer systems confidently. We acknowledge previous work which concluded that, in complex systems, failures ("accidents") may be inevitable and unavoidable. To further explore potential tools for increasing our confidence in complex systems, we review research in the field of Complexity Theory to seek potentially useful approaches and measures and find ourselves particularly interested in the potential usefulness of relationships between the magnitudes of events and their frequency of occurrence. Complexity Theory is found to have characterized naturally occurring systems and to potentially be the source of profitable application to the systems engineering challenge, viz., the creation of complex engineered systems. We are left with the tentative conclusion that truly complex systems, with our present understanding of complex behavior, cannot be designed with a degree of confidence that is acceptable given our current expectations. We recommend that the discipline of systems engineering must investigate this issue as a matter of priority and urgency and seek to develop approaches to respond to the challenge.